JP2005224901A - Edge detector - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an edge detector simply and highly precisely detecting the edge position of a shield material positioned in an optical path from an output of a line sensor and simplifying a light source. <P>SOLUTION: This edge detector is provided with a line sensor 1, a point light source 2 projecting a monochrome light having a flare angle toward the line sensor reaching its whole light receiving width, and a computing part analyzing the output of the line sensor to find the edge position of the shield material 7 positioned in the optical path of the monochrome light by; magnifies and detects an optical diffraction pattern formed in a shadow and the edge part of a detection object (shield material) positioned in the optical path; and simultaneously simplifies the constitution of the light source. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an edge detection device that can easily and accurately detect an edge position of a shield positioned in an optical path from an output of a line sensor.

  Recently, the printed circuit board has been multi-layered with high density mounting, and through holes are formed in the printed circuit board to electrically connect a plurality of layers. Such a through hole is exclusively formed by using a fine drill blade having a diameter of about 50 to 100 μm, for example, and rotating the drill blade at a high speed to form a hole of a predetermined depth in the printed circuit board. At this time, it is of course possible to select and use a drill blade of a predetermined diameter, attach this drill blade to the chuck of the drill without any runout, and further accurately grasp the tip position of the drill blade to determine the predetermined diameter. It is important to drill holes to the depth of. However, it is generally very difficult to mechanically measure the diameter (drill diameter) of this kind of small-diameter drill blade and to check the mounting state on the chuck, etc. A typical measuring means is used (for example, see Patent Documents 1, 2, and 3).

  However, the optical measuring method of a drill blade as shown in Patent Documents 1, 2, and 3 only measures the light shielding width by a line sensor or the like using light shielding of the drill blade, and has a diameter. However, it is difficult to accurately measure the diameter of a drill blade having a small diameter of 200 μm or less. That is, monochromatic parallel light such as laser light is exclusively used as the light source for this type of measurement. However, since diffraction of light occurs at the edge portion shielded by the drill blade, there is a problem that it is difficult to accurately measure the diameter of the drill blade due to the influence of this diffraction.

On the other hand, the present inventor uses the approximate expression obtained by approximating the diffraction pattern (intensity distribution) of the light previously generated by Fresnel diffraction using the hyperbolic second function sech (x) to easily and highly increase the edge position. A method for obtaining accuracy was proposed (see, for example, Patent Document 4).
JP 2003-170335 A Japanese Patent Application Laid-Open No. 7-306020 JP 7-260425 A Japanese Patent Application No. 2002-345958

  Incidentally, in the optical measurement methods shown in Patent Documents 1, 2, 3, and 4, monochromatic parallel light such as laser light is exclusively used as the light source. Specifically, laser light emitted from a laser diode (LD) is converted into a parallel light beam via a light projection lens (collimator lens) and projected onto the light receiving surface of the line sensor. The line sensor detects the shadow of the detection object (shielding object) positioned in the optical path and the Fresnel diffraction light pattern generated at the edge of the detection object (shielding object). Yes.

  However, since the light source for projecting monochromatic parallel light requires an optical element such as the above-described light projecting lens (collimator lens), the configuration becomes large and the manufacturing cost is increased. In addition, when trying to widen the beam bundle width of monochromatic parallel light, for example, a large-diameter projection lens (collimator lens) is required, and generally the optical distance between the LD and the projection lens can be set long. Since this is necessary, there is a problem such as an increase in the size of the light source.

  The present invention has been made in view of such circumstances, and its purpose is to determine the edge position of the shield positioned in the optical path from the output of the line sensor without incurring the complexity of the configuration of the light source. An object of the present invention is to provide an edge detection device that can be simply and accurately detected.

  In order to achieve the above-described object, the present invention is based on the spread of the laser light (monochromatic light) when the laser light (monochromatic light) emitted from the LD is projected as it is without passing through the projection lens (collimator lens). The shadow of the detection object (shield) positioned in the optical path is enlarged and projected onto the light receiving surface of the line sensor, and strictly speaking, no Fresnel diffraction of light occurs at the edge portion. The focus is on the fact that the light pattern projected onto the light receiving surface of the sensor can be substantially regarded as a Fresnel diffraction pattern.

Therefore, the edge detection apparatus according to the present invention is
<a> a line sensor in which a plurality of light receiving cells are arranged at a predetermined pitch in one direction;
<b> A point light source that projects monochromatic light having a divergence angle that reaches the entire light receiving width of the line sensor toward the plurality of light receiving cells of the line sensor;
<c> An arithmetic unit that obtains an edge position of the shielding object positioned in the optical path of the monochromatic light by analyzing the output of the line sensor, thereby detecting a detection object (shielding object) positioned in the optical path. ) And the light diffraction pattern generated at the edge of the detection object (shielding object) are enlarged and detected, and at the same time, the configuration of the light source is simplified.

  Preferably, as described in claim 2, the arithmetic unit calculates the optical system enlargement ratio SD according to the distance SD between the point light source and the line sensor and the distance WD between the shield and the line sensor. / (SD-WD) is configured to obtain the edge position of the shielding object by correcting the edge position of the light receiving pattern obtained by analyzing the output of the line sensor.

In particular, as described in claim 3, the shield is a round bar, and when the width of the shadow of the round bar is obtained from the output of the line sensor,
The calculation unit calculates the shadow width 2a of the round bar-like body obtained on the line sensor, the distance SD between the point light source and the line sensor, the distance WD between the shield and the line sensor, and the above The diameter 2r of the round bar is corrected based on the shadow width 2a, and 2r = 2a (SD−WD) / {(2a) ² + SD ² } ^1/2
It is characterized by asking.

  Note that the calculation unit may detect the edge position of the shadow generated on the light receiving surface of the line sensor, assuming that Fresnel diffraction occurs at the edge of the shielding unit, for example.

  According to the edge detection apparatus having the above-described configuration, it is only necessary to use a point light source that projects monochromatic light having a divergence angle that reaches the entire light receiving width of the line sensor as a light source. The monochromatic light (laser light) emitted from the laser diode (LD) may be projected as it is toward the line sensor without being used, and the configuration can be greatly simplified and the manufacturing cost can be reduced. Further, since the shadow of the detection object (shielding object) positioned in the optical path is enlarged and projected onto the line sensor, the detection accuracy of the shadow edge position can be increased by the enlargement ratio.

  Therefore, it is possible to achieve a great practical effect such that the detection accuracy can be improved while simplifying the overall configuration of the apparatus (particularly the configuration of the light source).

Hereinafter, an edge detection apparatus according to an embodiment of the present invention will be described with reference to the drawings, taking as an example an apparatus for measuring a diameter of a very small drill blade (drill diameter).
FIG. 1 shows a schematic configuration of a main part of an edge detection apparatus used for the diameter measurement. The edge detection device is generally provided so as to be opposed to a line sensor (light receiving unit) 1 including a plurality of light receiving cells arranged at a predetermined pitch w in one direction, and a light receiving surface of the line sensor 1. A light projecting unit 2 that projects monochromatic light 4 toward the entire width of the plurality of light receiving cells 1a of the line sensor 1 is provided. The line sensor 1 includes, for example, 102 light receiving cells arranged at a pitch of 85 μm and a light receiving surface having a long side of 8.7 mm and a short side of 0.08 mm. In addition, the apparatus main body 3 realized by a microcomputer or the like analyzes the output of the line sensor 1 (the amount of light received by each light receiving cell) and is positioned in the optical path of the monochromatic light 4 to be a shield (detection target). 7 plays a role of detecting the edge position in the arrangement direction of the light receiving cell 1a with high accuracy.

  Incidentally, the light projecting unit 2 includes a point light source 2a such as a laser diode (LD) that projects monochromatic light (laser light) having a predetermined divergence angle toward the line sensor 1 as it is. Such a light projecting unit 2 is integrated into a U-shaped housing 5 having a predetermined gap (space part) L together with the above-described line sensor 1 so as to face each other with the gap L interposed therebetween. And formed as one sensing unit. When a laser diode (LD) is used as the point light source 2a, the LD emits laser light having an elliptical intensity distribution so that the long axis direction of the laser light is the longitudinal direction of the line sensor 1. It is desirable to set the optical system.

  The apparatus body 3 includes an input buffer 3a that takes in the output of the line sensor 1 (the amount of light received by each light receiving cell) and obtains the light intensity distribution on the light receiving surface of the line sensor 1. In particular, the apparatus main body 3 receives monochromatic light having a predetermined light bundle width projected from the light projecting unit 2 in advance as the initial setting process by the line sensor 1, and based on the light intensity distribution at this time, A diffraction pattern detection unit 3b is provided for obtaining a diffraction pattern of monochromatic light projected by the light projecting unit 2 and obtaining a normalization parameter for the amount of light received by each light receiving cell 1a according to the reciprocal of the diffraction pattern as will be described later.

  Further, the apparatus main body 3 includes a normalizing means 3c for normalizing the output of the line sensor 1 according to the normalization parameter obtained by the diffraction pattern detecting means 3b, and the line sensor normalized by the normalizing means 3c. 1, the position of the end (edge) of the shielding object (detection target) 7, specifically, the edge position of the light receiving pattern of the line sensor 1 in the array direction of the light receiving cells is detected. And an edge detector 3b. As will be described later, this edge detection unit 3b assumes that Fresnel diffraction of monochromatic light occurs at the end (edge) of the shield (detection target) 7 and that the diffraction pattern is projected onto the light receiving surface of the line sensor 1. The edge position is obtained by analyzing the light intensity distribution of the diffraction pattern. In this edge detection apparatus, the apparatus main body 3 further uses a drill diameter measuring section 3e for measuring the diameter of a drill blade (drill diameter) as an inspection object by using the output of the edge detection section 3d. It is configured to include a runout detector 3f that detects the runout of the blade and a tip position detector 3g that detects the tip position of the drill blade.

  Here, the edge position detection processing in the edge detection unit 3d will be described. The edge detection unit 3d basically generates Fresnel diffraction at its end (edge) when a part of monochromatic parallel light is blocked by the shielding object (detection object) 7, and causes Fresnel diffraction. Paying attention to the distribution characteristic that the intensity of light reaching the light receiving surface of the line sensor 1 rises steeply in the vicinity of the edge position as described below and converges while oscillating as the distance from the edge position increases. According to the light intensity distribution on the light receiving surface of the line sensor 1, the position of the end (edge) of the shield 7 is detected with high accuracy.

That is, when the shielding object 7 has a plate shape that blocks the optical path of the monochromatic light 4 from one end side of the line sensor 1, the light intensity distribution due to Fresnel diffraction at the edge of the shielding object 7 is as shown in FIG. It rises sharply near the edge position and converges while vibrating as it moves away from the edge position. Such light intensity distribution characteristics are as follows: the wavelength of monochromatic parallel light is λ [nm], the distance from the edge of the inspection object (shield 7) to the light receiving surface is z [mm], and the edge position on the light receiving surface Assuming that x [μm] is [0], ∫ is an operation symbol indicating an integration from [x = 0] to [(2 / λz) ^1/2 · x]. Light intensity = (1/2) {[ 1/2 + S (x)] ² + [1/2 + C (x)] ² }
S (x) = ∫sin (π / 2) · U ² dU
C (x) = ∫cos (π / 2) · U ² dU
Represented as: However, U is a temporary variable.

As for the functions S (x) and C (x) in the above equation, the Fresnel function is used exclusively as shown in the mathematical formulas, and S (x) ′ ≈ (1/2) − where x is large. (1 / πx) cos (πx 2/2)
C (x) '≒ (1/2 ) + (1 / πx) sin (πx 2/2)
Can be approximated respectively. Therefore, basically, by using the approximate expressions S (x) ′ and C (x) ′, the edge position described above can be calculated from the received light intensity of each light receiving cell of the line sensor 1.

  However, when actually calculated, as shown in FIG. 3, the functions S (x) and C (x) and the approximate expressions S (x) ′ and C (x) ′ are converged after the rise ( Although it approximates very well in the second and subsequent peaks, it cannot be denied that there is a large shift in the first rising portion (first mountain). In particular, the characteristic of the first rising portion plays an important role in edge detection, and the deviation of the characteristic causes a decrease in the detection accuracy of the edge position.

Therefore, the present inventor previously described in Patent Document 4 that the first rising portion of the light intensity distribution on the light receiving surface by Fresnel diffraction of monochromatic parallel light, particularly the distribution characteristics of the first mountain, a, b, c respectively. As coefficient y = a · sech (bx + c)
It is found that the hyperbolic second function sech (x) approximates very well, and the output (light intensity) of the line sensor is analyzed using the hyperbolic second function sech (x) on the light receiving surface by the Fresnel diffraction. It was proposed that the position xo where the light intensity (relative value) is 0.25 in the light intensity distribution is detected as the edge position of the shield 7 in the arrangement direction of the light receiving cells.

Specifically, when the above-described hyperbolic second function is applied to the above-described formula of the light intensity distribution by Fresnel diffraction and approximated to the first rising portion (first mountain) of the light intensity, the hyperbolic second function sech ( x) is the light intensity = 1.37 sech {1.98 (2 / λz) ^1/2 x−2.39}
As shown. This approximate expression agrees with the theoretical expression of the light intensity distribution with an accuracy of about three digits. Where λ is the wavelength of light [nm], z is the distance from the edge to the light receiving surface [mm], and x is the coordinate [μm] where the edge position on the light receiving surface is [0]. The above coefficients are set under various units.

The algorithm of edge position detection processing using such a hyperbolic second function sech (x) will be described. When an inverse function of light intensity approximated using the hyperbolic second function sech (x) is calculated,
Y = (y / 1.37)
X = 1.98 (2 / λz) ^1/2 x
Anyway,
X = 2.39−ln {[1+ (1−Y ² ) ^1/2 ] / Y}
Can be expressed as

  Therefore, in the edge detection unit 3d described above, each normalized received light intensity y1, y2, which is first obtained from a plurality (m) of light receiving cells 1a in the line sensor 1 according to the procedure shown in FIG. To ym, a light receiving cell Cn that obtains a light receiving intensity larger than the above-mentioned reference light intensity [0.25] and a light receiving cell that obtains a light receiving intensity smaller than the above-mentioned reference light intensity [0.25]. Cn-1 is obtained (step S1). That is, two adjacent light receiving cells Cn and Cn-1 having a light receiving intensity of [0.25] in each of the plurality of light receiving cells 1a (C1, C2, to Cm) are obtained. Then, the received light intensity yn, yn-1 of each of the light receiving cells Cn, Cn-1 is divided by the coefficient [1.37] described above to be converted into light intensity Yn, Yn-1 on the XY coordinates ( Step S2).

Thereafter, the positions Xn and Xn-1 on the light receiving surface of the light receiving cells Cn and Cn-1 from which the light receiving intensities Yn and Yn-1 of the light receiving cells Cn and Cn-1 are obtained are expressed by the above-described approximate expression. Xn = 2.39-ln accordance ^{{[1+ (1-Yn 2} ) 1/2] / Yn}
Xn-1 = 2.39-ln {[1+ (1-Yn-1 ² ) ^1/2 ] / Yn-1}
As shown in FIG. 5, the relative position on the X axis is calculated by inverse transformation (light receiving position calculating means; step S3), and the position of the light receiving cell Cn and the light receiving position are shown in FIG. The difference Δx from the edge position where the intensity is [0.25] is expressed as Δx = W · [Xn / (Xn−Xn−1)]
(Interpolation calculation means; step S4). The difference Δx is the distance from the edge position xo where the light receiving intensity is [0.25] to the position of the light receiving cell Cn, and is thus measured from the first light receiving cell C1 on the entire light receiving surface of the line sensor 1. When the absolute position x of the edge is defined as n, where n is the cell number of the light receiving cell 1a from which the light quantity Y2 is obtained, and W is the arrangement pitch of the light receiving cells 1a,
Can be obtained as The relative positions Xn and Xn-1 obtained in the inverse transformation are
X = 1.98 (2 / λz) ^1/2 x
As shown, the value is multiplied by [1.98 (2 / λz) ^1/2 ], but this term is substantially eliminated by taking the ratio in the above-described interpolation calculation.

  This interpolation calculation may be executed using the above-described approximate expression. However, if the change in light intensity between the two light receiving cells Cn and Cn-1 can be regarded as linear, it is simple. Simple linear interpolation may be used. Here, a position where the light intensity is [0.25] is found between adjacent light receiving cells 1a, and two light receiving cells Cn and Cn-1 having the position as a cell boundary are specified, but the above position is simply sandwiched. Two or more light receiving cells may be specified. However, in this case, it is only necessary to prevent the calculation accuracy from being lowered by performing the interpolation calculation using the approximate expression described above. Further, the inverse transformation described above can be executed instantaneously, for example, by using a table in which the calculated values are stored in advance, thereby greatly reducing the calculation processing load.

The difference between the relative positions Xn, Xn-1 on the light receiving surface of the light receiving cells Cn, Cn-1 and the position (edge position) xo where the light receiving intensity is [0.25] and the position of the light receiving cell Cn. Paying attention to Δx, the light receiving intensity at the light receiving cell Cn, and the wavelength λ of the monochromatic parallel light, the distance between the edge of the shield 7 and the light receiving surface of the line sensor 1 from the hyperbolic second function sech (x), That is, the distance z in the optical path direction can also be obtained (step S5). Specifically, this distance calculation is basically performed by using the above-described formula approximating the above-mentioned Fresnel diffraction at the first peak. Light intensity A (x) = 1.37 · sech {1.98 (2 / λz) ^1/2 x-2.39}
Solve for the distance z from
z = (2 / λ) {1.98 · x / [arcsech (A (x) /1.37) +2.39]} ²
As described above, the distance z between the edge of the shield 7 and the light receiving surface of the line sensor 1 can be calculated.

In this case, when the edge position in the arrangement direction of the light receiving cells described above is obtained, the position of the light receiving cell Cn where the light intensity is higher than [0.25] is used, and this position and the edge position are determined. From the difference Δx,
z = (2 / λ) {1.98 · Δx / [arcsech (yn / 1.37) +2.39]} ²
As a result, the distance z between the edge of the shield 7 and the light receiving surface of the line sensor 1 can be easily obtained. Especially denominator term in the above equation, the aforementioned Xn = 2.39-ln {[1+ (1-Yn 2) 1/2] / Yn}
Therefore, the above calculation is performed by z = (2 / λ) {1.98 · Δx / Xn} ²
It becomes possible to calculate more simply as follows.

  By the way, when the shielding object 7 is a small-diameter rod-like body as described above, for example, a drill blade, Fresnel diffraction of the monochromatic parallel light 4 occurs at both side portions of the drill blade, so the Fresnel diffraction pattern on the light receiving surface of the line sensor 1 is For example, as shown in FIG. 6 (a), the pattern has a symmetry that converges while vibrating on both sides of the center position of the drill blade, and the light receiving intensity at each light receiving cell of the line sensor 1 is as shown in FIG. As shown in b). Moreover, when the drill diameter is 200 μm or less, the received light intensity may not decrease to [0.25]. Therefore, as described above, it is impossible to accurately obtain the position where the light amount is [0.25] using the approximate expression of Fresnel diffraction.

However, the diffraction pattern shown in FIG. 6 (a) is considered to be a combination of Fresnel diffractions generated on both sides of the shield (drilling blade) 7 approximately as shown in FIG. 6 (c). be able to. Therefore, for example, the light intensity A (x) that has passed through the shield (drilling blade) 7 having the radius r is the light intensity A (x) L of the left diffraction pattern and the light intensity A (x) of the right diffraction pattern. ) R and A (x) = A (x) L + A (x) R
= 1.37 sech {-1.98 (2 / λz) ^1/2 (x + r) -2.39}
+1.37 sech {1.98 (2 / λz) ^1/2 (x−r) −2.39}
Can be understood as However, as the drill diameter becomes smaller, the overlap in the vicinity of [0.25] in the light intensities A (x) L and A (x) R of the left and right diffraction patterns greatly affects the light receiving surface of the line sensor 1. Therefore, the position where the light amount is [0.25] cannot be directly detected as the edge position as described above.

Therefore, in the drill diameter measuring unit 3e, the light intensity A (x) L, A (x) R of the left and right diffraction patterns described above is almost affected by the other diffraction pattern at the first rising portion. Particular attention is paid to a region where the light intensity (light quantity) is [0.5 to 0.9]. For example, the light intensity (light quantity) is [0. 75], the rough edge positions X _R and X _L are obtained (steps S11 and S12). Then, a light shielding width 2a having a light quantity of [0.75] in the diffraction pattern A (x) is obtained from these left and right rough edge positions X _R and X _L (step S13), and the above-described operation is performed according to the light shielding width 2a. The drill diameter is obtained by calculating back the radius r of the drill blade (step S14).

Specifically, the edge position X _R where the light intensity is [0.75] is obtained from the right diffraction pattern A (x) _R as follows. That is, the light intensity y
y = 1.37 sech {1.98 (2 / λz) ^1/2 X−1.21}
In
Y = y / 1.37
And put
sech ⁻¹ (Y) = ± ln [{1+ (1−Y ² ) ^1/2 } / Y]
= X'-1.21
However, 0 <y ≦ 1.37, 0 <Y ≦ 1.0, X ′ = 1.98 (2 / λz) ^1/2 X
It becomes.

Therefore, the measured value (normalized data y0, y1, y2,..., Y101) in each light receiving cell of the line sensor 1 composed of 102 cells is between the [n−1] th cell and the nth cell. Where the light intensity is [0.75], and the light intensity in the [n-1] -th and n-th cells is yn-1, yn,
Yn-1 = yn-1 / 1.37, Yn = yn / 1.37
As in the case shown in FIG. 5, the position where the light intensity is [0.75] is expressed as X′n−1 = 1.21−ln [{1+ (1−Yn−1 ² ) ^1/2 } / Yn-1]
X′n = 1.21−ln [{1+ (1−Yn ² ) ^1/2 } / Yn]
Can be mapped respectively. As a result, the edge position X _R is changed from these mapping positions to X _R [μm] = w {n−X′n / (X′n−X′n−1)}.
Can be obtained easily and accurately by interpolation processing. Where w is the width of the cell. As described above, X′n and X′n−1 are not original cell positions but are values multiplied by 1.98 (2 / λz) ^1/2. Since this term is substantially eliminated by obtaining, the interpolation process can be performed accurately even if the distance z is unknown.

Similarly, an edge position X _{L at} which the light amount is [0.75] is obtained from the left diffraction pattern A (x) L. Then, according to the edge positions X _R and X _L obtained from these diffraction patterns A (x) R and A (x) L, the light shielding width 2a at the position where the light quantity is [0.75] is set to 2a = X _R − X _L
Asking.

Next, the above light quantity [0.75] and half the value a of the light-shielding width 2a are added to the equation representing the diffraction pattern obtained by combining the light intensities A (x) R and A (x) L of the right and left diffraction patterns. Substituting and calculating the radius r of the drill blade in reverse. For the reverse calculation of r, for example, numerical calculation may be performed using the Newton method.
Specifically, f (r) = 1.37 sech {-1.98 (2 / λz) ^1/2 (a + r) -2.39}
+1.37 sech {1.98 (2 / λz) ^1/2 (ar) -2.39} −0.75
Then, the derivative is f ′ (r) = − 2.71 (2 / λz) ^1/2
Xsech {-1.98 (2 / λz) ^1/2 (a + r) -2.39}
Xtanh {-1.98 (2 / λz) ^1/2 (a + r) -2.39}
-2.71 (2 / λz) ^1/2
× sech {1.98 (2 / λz) ^1/2 (ar) -2.39}
X tanh {1.98 (2 / λz) ^1/2 (ar) -2.39}
It becomes.

Therefore, the initial value r0 of r is set to r0 = {2a−0.845 (λz) ^1/2 } / 2
age,
r _n = r _n−1 −f (r _n−1 ) / f ′ (r _n−1 )
n = 1,2,3, ...
If it is repeatedly calculated until the absolute value of [r _n −r _n−1 ] is, for example, 0.01 or less, the converged r can be obtained as the radius of the drill. Therefore, the drill diameter can be obtained as twice the radius r, specifically as 2r.

  As for the drill diameter (radius r) calculated in this way, when the distance z between the drill blade and the line sensor 1 is known in advance, for example, as shown in FIG. 8, the light shielding width 2a and the diameter 2r. As a relationship, the table may be stored in advance. If such a table 3h is used, since the back-calculation process using the Newton method mentioned above becomes unnecessary each time, it becomes possible to measure a drill diameter easily.

When the drill diameter is somewhat thick and the interference between the light intensities A (x) R and A (x) L of the right and left diffraction patterns is negligible, the light intensity A (x ) R, A (x) L, for example, 0.75 = 1.37 sech {1.98 (2 / λz) ^1/2 (ar) -2.39}
Just solve
2r = 2a-0.845 (λz) ^1/2
The radius r can be obtained as follows. That is, the diameter (drill diameter) 2r of the drill blade can be easily obtained by simply subtracting [0.845 (λz) ^1/2 ], which is the specified value of the optical system, from the light shielding width 2a at the light amount [0.75]. Can do.

  The edge detection apparatus according to the present invention is configured to project monochromatic light emitted from the point light source 2 toward the line sensor 1 as it is, as described above. The monochromatic light projected from the light source 2 toward the line sensor 1 has a predetermined divergence angle and is not parallel light. Therefore, in a strict sense, the end (edge) of the shielding object (detection target) 7 is used. In Fresnel diffraction does not occur. However, since the divergence angle of the laser beam emitted from the LD is generally as narrow as about 17 °, this is regarded as parallel light, that is, considering that Fresnel diffraction occurs, as described above. Even if edge detection is performed, almost no measurement error occurs.

  When measuring the diameter of a drill blade, the drill blade (detection target) 7 is usually set at a predetermined position in the optical path of monochromatic light and the measurement is performed. The distance z between 7 and the line sensor 1 does not become a problem. In other words, the diameter z can be measured by giving a fixed distance z between the drill blade (detection target) 7 and the line sensor 1. Then, the detection optical system in this edge detection apparatus enlarges the shadow of monochromatic light (laser light) emitted from the point light source 2 and projects it onto the light receiving surface of the line sensor 1 as schematically shown in FIG. It can be expressed as a magnifying optical system. The shadow (drill diameter) by the drill blade (detection target) 7 is determined by the ratio of the distance SD between the point light source 2 and the line sensor 1 and the distance WD between the drill blade (detection target) 7 and the line sensor 1. The image is projected onto the line sensor 1 at a fixed magnification.

  Specifically, the shadow of the drill blade 7 when monochromatic parallel light is used corresponds to the diameter of the drill blade 7 and is projected onto the line sensor 1, for example, as shown in FIG. When monochromatic light having a predetermined divergence angle from 2 is used, the shadow of the drill blade 7 is enlarged and projected onto the line sensor 1 as shown in FIG. Therefore, if the shadow width 2a of the enlarged drill blade 7 is obtained as described above, the true drill diameter can be obtained by increasing the measurement accuracy from the measured value (shadow width) 2a by the enlargement ratio. It becomes possible.

That is, when the diameter of the drill blade 7 is 2r, according to the optical system described above, the shadow width a of the drill blade 7 is 2a = 2r · SD / (SD−WD).
And can be detected. Therefore, if the shadow width 2a of the drill blade 7 is measured as described above, the true diameter of the drill blade 7 is calculated back from the enlargement ratio, and 2r = 2a (SD-WD) / SD.
At the same time, and at the same time, the measurement error can be reduced by [(SD−WD) / SD] to increase the measurement accuracy.

Strictly speaking, since the drill blade 7 has a substantially round bar shape, the edge position with respect to monochromatic light is a position on a tangent line passing through the point light source 2 as shown in FIG. Therefore, the shadow width 2 a obtained on the line sensor 1 is slightly narrower than the actual diameter of the drill blade 7. However, since the distance SD between the point light source 2 and the line sensor 1 and the distance WD between the drill blade (detection target) 7 and the line sensor 1 are clear, respectively.
2r = 2a (SD-WD) / {(2a) ² + SD ² } ^1/2
As a result, the true diameter 2r of the drill blade 7 can be easily calculated.

  The following table shows the experimental results obtained by measuring wires having different diameters as described above in an optical system with SD of 50 mm and WD of 25 mm.

  As shown in this experimental example, it is confirmed that the drill diameter can be obtained with sufficiently high measurement accuracy even if the diameter of the drill blade 7 is measured by considering the above-described magnifying optical system as a Fresnel diffraction pattern. did it. In addition, in this edge detection device, the monochromatic light (laser light) emitted from the LD is simply projected onto the line sensor 1 as it is, and the monochromatic parallel light using a light projecting lens (collimator lens) as in the prior art. Therefore, the configuration of the light source 2 can be greatly simplified, and the component cost can be greatly reduced. In addition, it is only necessary to perform optical adjustment between the LD and the line sensor 1, and no special alignment accuracy is required. For example, the LD and the line sensor 1 are practically accurate enough to be mounted on a printed circuit board. In addition, an edge detection apparatus having sufficient measurement accuracy can be realized.

  Further, according to the edge detection device having the above-described configuration, when detecting the runout of the drill blade 7, the runout can be detected in an enlarged manner. Can be detected quickly and with high sensitivity. If the drill diameter is large, it is expected that the shadow will not enter the light receiving width of the line sensor 1. In this case, two line sensors 1 may be provided side by side in the longitudinal direction, and the installation interval of these line sensors 1 may be measured in advance at the time of shipping inspection or the like. Then, one side of the shadow of the drill blade 7 is measured using the two line sensors 1, and the shadow width 2a of the drill blade 7 is obtained from each edge measurement position and the installation interval of the line sensor 1. It ’s fine. When the diameter of the drill blade 7 is thin, the axial center position is shifted to one line sensor side, and the shadow width 2a of the drill blade 7 is measured using only this one line sensor 1. . It is also possible to use only one line sensor 1 and measure the drill diameter while changing the enlargement ratio. Specifically, the distance WD between the drill blade (detection target) 7 and the line sensor 1 is lengthened (closer to the projector 2) for those with a small drill diameter, and conversely the distance WD for those with a large drill diameter. The shadow width 2a may be measured by shortening (closer to the line sensor 1).

  The present invention is not limited to the embodiment described above. For example, the distance SD between the point light source 2 and the line sensor 1 and the distance WD between the drill blade (detection target) 7 and the line sensor 1 are determined in advance according to specifications such as the diameter width of the drill blade to be measured. It ’s fine. In addition, here, the case of measuring the diameter of the drill blade has been described as an example, but the same applies to the case of measuring the diameter of various wire rods, the case of detecting the edge position of a sheet-like object positioned at a predetermined measurement position, etc. Can be applied. However, when the distance WD between the detection object 7 and the line sensor 1 changes, the enlargement ratio itself also changes. Therefore, it is not preferable to employ the above-described enlargement optical system from the point light source, and the monochrome parallelism. Light should be used. In addition, the present invention can be variously modified and implemented without departing from the scope of the invention.

The figure which shows the principal part schematic structure of the edge detection apparatus which concerns on one Embodiment of this invention. The figure which shows the intensity | strength pattern of the light which produced the Fresnel diffraction by the edge of the shield. The figure which contrasts the theoretical value of the light intensity distribution by a Fresnel diffraction, and the approximate characteristic using a function. The figure which shows an example of the procedure of the edge detection process from a Fresnel diffraction pattern. The figure which shows the processing concept of the edge detection shown in FIG. The figure for demonstrating the measurement principle of the diffraction pattern and edge diameter which arise with a micro diameter drill blade. The figure which shows the example of the measurement processing procedure of edge diameter. The figure which shows the structural example of the table which shows the relationship between the light-shielding width in the position of light quantity [0.75], and the radius of a drill. The figure which shows typically the optical system of the edge detection apparatus which concerns on this invention. The figure which shows the light intensity pattern of the shadow of a drill blade at the time of using monochromatic parallel light, and the light intensity pattern of the shadow of a drill blade at the time of using the monochromatic light which has a predetermined divergence angle.

Explanation of symbols

1 Line sensor 2 Light emitter (light source)
3 Device body 3a Diffraction pattern detection unit 3d Edge detection unit 3e Drill diameter measurement unit 3h Table 7 Shield (drill blade)

Claims (4)

A line sensor in which a plurality of light receiving cells are arranged at a predetermined pitch in one direction;
A point light source that projects monochromatic light having a divergence angle that reaches the entire light receiving width of the line sensor toward the plurality of light receiving cells of the line sensor;
An edge detection apparatus comprising: an arithmetic unit that obtains an edge position of the shielding object positioned in the optical path of the monochromatic light by analyzing an output of the line sensor.   The arithmetic unit is configured to detect the line sensor according to a distance SD between the point light source and the line sensor and a magnification ratio SD / (SD−WD) of an optical system determined according to a distance WD between the shielding object and the line sensor. The edge detection apparatus according to claim 1, wherein the edge position of the light-shielding pattern obtained by analyzing the output is corrected to obtain the edge position of the shielding object. When the shielding object is a round bar-like body, and the shadow width of the round bar-like body is obtained from the output of the line sensor, the calculation unit calculates the shadow width 2a of the round bar-like body obtained on the line sensor, Based on the distance SD between the point light source and the line sensor, the distance WD between the shielding object and the line sensor, and the shadow width 2a, the diameter 2r of the round bar-like body is set to 2r = 2a (SD− WD) / {(2a) ² + SD ² } ^1/2
The edge detection device according to claim 1, which is obtained as follows.   2. The edge detection according to claim 1, wherein the calculation unit is configured to detect an edge position of a shadow generated on a light receiving surface of the line sensor on the assumption that Fresnel diffraction is generated at an edge of the shielding unit. apparatus.

Images